Hasil untuk "Greek philology and language"

Philip Grawe

Matheus Belarmino, Rackel Coelho, Roberto Lotudo et al.

Large Language Models (LLMs) have been increasingly used to optimize the analysis and synthesis of legal documents, enabling the automation of tasks such as summarization, classification, and retrieval of legal information. This study aims to conduct a systematic literature review to identify the state of the art in prompt engineering applied to LLMs in the legal context. The results indicate that models such as GPT-4, BERT, Llama 2, and Legal-Pegasus are widely employed in the legal field, and techniques such as Few-shot Learning, Zero-shot Learning, and Chain-of-Thought prompting have proven effective in improving the interpretation of legal texts. However, challenges such as biases in models and hallucinations still hinder their large-scale implementation. It is concluded that, despite the great potential of LLMs for the legal field, there is a need to improve prompt engineering strategies to ensure greater accuracy and reliability in the generated results.

Stergios Chatzikyriakidis, Dimitris Papadakis, Sevasti-Ioanna Papaioannou et al.

We present an extended Greek Dialectal Dataset (GRDD+) 1that complements the existing GRDD dataset with more data from Cretan, Cypriot, Pontic and Northern Greek, while we add six new varieties: Greco-Corsican, Griko (Southern Italian Greek), Maniot, Heptanesian, Tsakonian, and Katharevusa Greek. The result is a dataset with total size 6,374,939 words and 10 varieties. This is the first dataset with such variation and size to date. We conduct a number of fine-tuning experiments to see the effect of good quality dialectal data on a number of LLMs. We fine-tune three model architectures (Llama-3-8B, Llama-3.1-8B, Krikri-8B) and compare the results to frontier models (Claude-3.7-Sonnet, Gemini-2.5, ChatGPT-5).

Brendon Boldt, David Mortensen

We introduce CSAR, an algorithm for inducing morphemes from emergent language corpora of parallel utterances and meanings. It is a greedy algorithm that (1) weights morphemes based on mutual information between forms and meanings, (2) selects the highest-weighted pair, (3) removes it from the corpus, and (4) repeats the process to induce further morphemes (i.e., Count, Select, Ablate, Repeat). The effectiveness of CSAR is first validated on procedurally generated datasets and compared against baselines for related tasks. Second, we validate CSAR's performance on human language data to show that the algorithm makes reasonable predictions in adjacent domains. Finally, we analyze a handful of emergent languages, quantifying linguistic characteristics like degree of synonymy and polysemy.

Qiao Wang, Adnan Labib, Robert Swier et al.

GenQuest is a generative text adventure game that leverages Large Language Models (LLMs) to facilitate second language learning through immersive, interactive storytelling. The system engages English as a Foreign Language (EFL) learners in a collaborative "choose-your-own-adventure" style narrative, dynamically generated in response to learner choices. Game mechanics such as branching decision points and story milestones are incorporated to maintain narrative coherence while allowing learner-driven plot development. Key pedagogical features include content generation tailored to each learner's proficiency level, and a vocabulary assistant that provides in-context explanations of learner-queried text strings, ranging from words and phrases to sentences. Findings from a pilot study with university EFL students in China indicate promising vocabulary gains and positive user perceptions. Also discussed are suggestions from participants regarding the narrative length and quality, and the request for multi-modal content such as illustrations.

Rudolf Henneböhl

Sayantan Adak, Daivik Agrawal, Animesh Mukherjee et al.

We investigate the knowledge of object affordances in pre-trained language models (LMs) and pre-trained Vision-Language models (VLMs). A growing body of literature shows that PTLMs fail inconsistently and non-intuitively, demonstrating a lack of reasoning and grounding. To take a first step toward quantifying the effect of grounding (or lack thereof), we curate a novel and comprehensive dataset of object affordances -- Text2Afford, characterized by 15 affordance classes. Unlike affordance datasets collected in vision and language domains, we annotate in-the-wild sentences with objects and affordances. Experimental results reveal that PTLMs exhibit limited reasoning abilities when it comes to uncommon object affordances. We also observe that pre-trained VLMs do not necessarily capture object affordances effectively. Through few-shot fine-tuning, we demonstrate improvement in affordance knowledge in PTLMs and VLMs. Our research contributes a novel dataset for language grounding tasks, and presents insights into LM capabilities, advancing the understanding of object affordances. Codes and data are available at https://github.com/sayantan11995/Text2Afford

Carl Edwards, Qingyun Wang, Lawrence Zhao et al.

Language-molecule models have emerged as an exciting direction for molecular discovery and understanding. However, training these models is challenging due to the scarcity of molecule-language pair datasets. At this point, datasets have been released which are 1) small and scraped from existing databases, 2) large but noisy and constructed by performing entity linking on the scientific literature, and 3) built by converting property prediction datasets to natural language using templates. In this document, we detail the $\textit{L+M-24}$ dataset, which has been created for the Language + Molecules Workshop shared task at ACL 2024. In particular, $\textit{L+M-24}$ is designed to focus on three key benefits of natural language in molecule design: compositionality, functionality, and abstraction.

Manuel de Buenaga, Francisco Javier Bueno

The GPT (Generative Pre-trained Transformer) language models are an artificial intelligence and natural language processing technology that enables automatic text generation. There is a growing interest in applying GPT language models to university teaching in various dimensions. From the perspective of innovation in student and teacher activities, they can provide support in understanding and generating content, problem-solving, as well as personalization and test correction, among others. From the dimension of internationalization, the misuse of these models represents a global problem that requires taking a series of common measures in universities from different geographical areas. In several countries, there has been a review of assessment tools to ensure that work is done by students and not by AI. To this end, we have conducted a detailed experiment in a representative subject of Computer Science such as Software Engineering, which has focused on evaluating the use of ChatGPT as an assistant in theory activities, exercises, and laboratory practices, assessing its potential use as a support tool for both students and teachers.

Zofia Malisz, Jan Foremski, Małgorzata Kul

We present a speech database and a phoneme-level language model of Polish. The database and model are designed for the analysis of prosodic and discourse factors and their impact on acoustic parameters in interaction with predictability effects. The database is also the first large, publicly available Polish speech corpus of excellent acoustic quality that can be used for phonetic analysis and training of multi-speaker speech technology systems. The speech in the database is processed in a pipeline that achieves a 90% degree of automation. It incorporates state-of-the-art, freely available tools enabling database expansion or adaptation to additional languages.

Renxi Wang, Haonan Li, Xudong Han et al.

Large language models (LLMs) have achieved success in acting as agents, which interact with environments through tools such as search engines. However, LLMs are optimized for language generation instead of tool use during training or alignment, limiting their effectiveness as agents. To resolve this problem, previous work has first collected interaction trajectories between LLMs and environments, using only trajectories that successfully finished the task to fine-tune smaller models, making fine-tuning data scarce and acquiring it both difficult and costly. Discarding failed trajectories also leads to significant wastage of data and resources and limits the possible optimization paths during fine-tuning. In this paper, we argue that unsuccessful trajectories offer valuable insights, and LLMs can learn from these trajectories through appropriate quality control and fine-tuning strategies. By simply adding a prefix or suffix that tells the model whether to generate a successful trajectory during training, we improve model performance by a large margin on mathematical reasoning, multi-hop question answering, and strategic question answering tasks. We further analyze the inference results and find that our method provides a better trade-off between valuable information and errors in unsuccessful trajectories. To our knowledge, we are the first to demonstrate the value of negative trajectories and their application in agent-tunning scenarios. Our findings offer guidance for developing better agent-tuning methods and low-resource data usage techniques.

Jiahao Huo, Yibo Yan, Boren Hu et al.

Projecting visual features into word embedding space has become a significant fusion strategy adopted by Multimodal Large Language Models (MLLMs). However, its internal mechanisms have yet to be explored. Inspired by multilingual research, we identify domain-specific neurons in multimodal large language models. Specifically, we investigate the distribution of domain-specific neurons and the mechanism of how MLLMs process features from diverse domains. Furthermore, we propose a three-stage mechanism for language model modules in MLLMs when handling projected image features, and verify this hypothesis using logit lens. Extensive experiments indicate that while current MLLMs exhibit Visual Question Answering (VQA) capability, they may not fully utilize domain-specific information. Manipulating domain-specific neurons properly will result in a 10% change of accuracy at most, shedding light on the development of cross-domain, all-encompassing MLLMs in the future. The source code is available at https://github.com/Z1zs/MMNeuron.

А.И. Солопов, Ильдар Равильевич Гимадеев

В статье разбираются цветообозначения, относящиеся к вину у Плиния Старшего (colores uinis quattuor: albus, fuluus, sanguineus, niger Plin. n. h. XIV 80). Вместо привычного нам белого и красного вина здесь перечисляются четыре цвета: albus, обозначающий «матовый белый», fuluus, отражающий широкий спектр между светло-коричневым и темно-желтым, sanguineus – цвет крови, niger – «черный блестящий». Данная классификация отчасти противоречит другим местам у Плиния, где он разделяет цвет вина на условно «красный» и «белый» (niger и candicāns) или сообщает, что у вина есть множество самых разных оттенков (n. h. XIV 15). Приводится сравнение с самой подробной античной классификацией цвета вина, разработанной для греческого языка у Афинея (μέλας, λευκός, κιῤῥός «желтовато-коричневый» (Athen. 32 d)) и у Галена (Galen. ad Hippocr. de acut. morb. uictu III 1), который прибавляет сюда цвета ξανθός и ἐρυθρός, т. е. «красный», ожидаемый нами. Для сопоставления авторы приводят также все возможные варианты цветообозначений, характеризующих вино у римских авторов, разделяя его на условные «белое» и «красное». Для объяснения некоторой путаницы в цветообозначении описываются технологические процессы, влияющие на цвет готового продукта. Авторы приходят к выводу о том, что Плиний в n. h. XIV 80 дает скорее бытовую классификацию цветообозначения вина. Он подходит к вопросу не как винодел, но как простой обыватель, не разделяя вино по цвету на привычные красное или черное и белое, но пытаясь передать его разнообразные оттенки как любитель. The article examines the color designation of wine by Pliny the Elder (colores uinis quattuor: albus, fuluus, sanguineus, niger Plin. n. h. XIV 80). Instead of the usual white and red wines, four colors are listed here: albus, meaning “matte white”, fuluus, reflecting a wide spectrum between light brown and dark yellow, sanguineus – the color of blood, niger – “shiny black”. This classification partly contradicts other places in Pliny, where he divides wine into conventionally “red” and “white” (niger and candicāns) or reports that wine has many different shades (n. h. XIV 15). A comparison is made with the most detailed ancient classification of wine color, developed for the Greek language by Athenaeus (μέλας, λευκός, κιῤῥός “orange tawny” (ap. Athen. 32 d)) and by Galen (Galen. ad Hippocr. de acut. morb. uictu III 1), which adds here the colors ξανθός and ἐρυθρός, i.e. “red”, which we expect. The authors also provide for comparison all possible options for the color designation of wine by Roman authors, dividing it into conventional “white” and “red”. To explain some of the confusion in the color designation of wine, some technological processes are described that affect the color of the finished product. The authors come to the conclusion that Pliny in n. h. XIV 80 gives a rather everyday classification of the color designation of wine. He approaches the issue not as a winemaker, but as a simple layman, not dividing wine by color into the usual red or black and white, but trying to convey its various shades as an amateur.

Shangshang Zheng, He Bai, Yizhe Zhang et al.

Large Language Models (LLMs) might hallucinate facts, while curated Knowledge Graph (KGs) are typically factually reliable especially with domain-specific knowledge. Measuring the alignment between KGs and LLMs can effectively probe the factualness and identify the knowledge blind spots of LLMs. However, verifying the LLMs over extensive KGs can be expensive. In this paper, we present KGLens, a Thompson-sampling-inspired framework aimed at effectively and efficiently measuring the alignment between KGs and LLMs. KGLens features a graph-guided question generator for converting KGs into natural language, along with a carefully designed importance sampling strategy based on parameterized KG structure to expedite KG traversal. Our simulation experiment compares the brute force method with KGLens under six different sampling methods, demonstrating that our approach achieves superior probing efficiency. Leveraging KGLens, we conducted in-depth analyses of the factual accuracy of ten LLMs across three large domain-specific KGs from Wikidata, composing over 19K edges, 700 relations, and 21K entities. Human evaluation results indicate that KGLens can assess LLMs with a level of accuracy nearly equivalent to that of human annotators, achieving 95.7% of the accuracy rate.

Mario Giulianelli, Andrey Kutuzov, Lidia Pivovarova

Morphological and syntactic changes in word usage (as captured, e.g., by grammatical profiles) have been shown to be good predictors of a word's meaning change. In this work, we explore whether large pre-trained contextualised language models, a common tool for lexical semantic change detection, are sensitive to such morphosyntactic changes. To this end, we first compare the performance of grammatical profiles against that of a multilingual neural language model (XLM-R) on 10 datasets, covering 7 languages, and then combine the two approaches in ensembles to assess their complementarity. Our results show that ensembling grammatical profiles with XLM-R improves semantic change detection performance for most datasets and languages. This indicates that language models do not fully cover the fine-grained morphological and syntactic signals that are explicitly represented in grammatical profiles. An interesting exception are the test sets where the time spans under analysis are much longer than the time gap between them (for example, century-long spans with a one-year gap between them). Morphosyntactic change is slow so grammatical profiles do not detect in such cases. In contrast, language models, thanks to their access to lexical information, are able to detect fast topical changes.

Kshitij Gupta

Large pre-trained language models have brought remarkable progress in NLP. Pre-training and Fine-tuning have given state-of-art performance across tasks in text processing. Data Augmentation techniques have also helped build state-of-art models on low or zero resource tasks. Many works in the past have attempted at learning a single massively-multilingual machine translation model for zero-shot translation. Although those translation models are producing correct translations, the main challenge is those models are producing the wrong languages for zero-shot translation. This work and its results indicate that prompt conditioned large models do not suffer from off-target language errors i.e. errors arising due to translation to wrong languages. We empirically demonstrate the effectiveness of self-supervised pre-training and data augmentation for zero-shot multi-lingual machine translation.

S. Luraghi

The encoding of the semantic role of direction may display animacy based differential marking. Cross-linguistic data also show that both human and inanimate direction may be encoded in the same way as the semantic role of recipient. After briefly surveying some attested patterns in the encoding of these three semantic roles, the paper concentrates on three Ancient Indo-European languages, Hittite, Latin and Ancient Greek. Among them, only Hittite makes use of the dative case to encode direction, while in the other languages the dative is limited to the role of recipient. Homeric Greek displays a cross-linguistically infrequent pattern, with the illative preposition extending to human direction. This pattern is dropped in Attic-Ionic prose.

Антонина В. Костић

One of the main goals of foreign language teaching is enabling effective communication through giving students tools to express their views and interact with other language speakers. There are many factors that affect the process of oral production development, as well as various strategies that can be applied in the process of motivating students to acquire knowledge and express themselves orally. Oratory techniques can be very useful in this process, because they can help the learner prepare for oral production in a foreign language. The aim of this paper is to provide an overview of rhetorical elements and oratory techniques that can be applied in the case of oral production in Modern Greek as a foreign language. The paper primarily compares the views expressed by Durbaba (2011) on the successful foreign language oral production and what Nusic (2009) and Avramovic (2008) underline as important for oratory. In addition, we present the oratory techniques that can be used in learning correct articulation and accentuation at A1/A2 (CEFR) level, as well as in preparation for a wellargued oral presentation on B1/B2 (CEFR) level of Modern Greek, with emphasis on learning Modern Greek as a foreign language on university level, at the Department of Modern Greek Studies of the Faculty of Philology, University of Belgrade. In addition, we point out the influence of affective factors, such as anxiety and motivation, on oral expression, as well as the ways in which oratory techniques can be a great tool for foreign language teachers in the process of motivating their audience (the learners).

Susanne Aretz

Andrew Cain

In the years leading up to his work on Paul, Jerome had become hardened in the conviction that biblical scholars should possess a mastery of the biblical languages, Hebrew and Greek, so that they can read Scripture in its original form. During his stay in Rome between 382 and 385, he had experimented with this back-to-the-sources approach in a number of shorter exegetical set pieces, but it was not until he embarked on his opus Paulinum that he was able finally to apply it systematically in the context of commentaries on whole biblical books. This chapter explores, through detailed case studies, how he develops his ad fontes methodology in the four Pauline commentaries and cumulatively builds the case for Hebrew and Greek philology being absolutely vital to serious study of the Bible, all the while attempting to demonstrate by example that he is the model biblical scholar.